You're right of course. I even do this pretty regularly. Dunno what I was 
thinking. 

Sent from my iPhone

On Aug 31, 2011, at 12:48 AM, Jonas Drewsen <[email protected]> wrote:

> On 31/08/11 01.17, Sean Kelly wrote:
>> On Aug 30, 2011, at 4:14 PM, Sean Kelly wrote:
>>> 
>>> HTTP really isn't the ideal protocol for this sort of thing.  By default, 
>>> the entire file would be in the message body, which isn't really 
>>> asynchronous.  Chunking is an alternative, but there's no way to do this 
>>> from the client-side, it's purely a server thing.
>> 
>> I stand corrected--it looks like libcurl lets you enable chunking.  I will 
>> say, however, that I've never actually see a chunked upload from a web 
>> browser, regardless of size.  It's kind of interesting that libcurl will do 
>> this.
> 
> Chunked transfer is only really necessary when you do not know the size in 
> advance. I doesn't mean that you cannot use asynchronous IO when doing 
> unchunked transfers. When you know the file size you can just set the file 
> size in the http header and send the file asynchronously piece by piece.
> 
> /Jonas

Reply via email to